Image-to-Text
Transformers
Safetensors
mistral3
text-generation
ocr
document-understanding
vision-language
pdf
tables
forms
Eval Results
🇪🇺 Region: EU
Instructions to use lightonai/LightOnOCR-1B-1025 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use lightonai/LightOnOCR-1B-1025 with Transformers:
# Use a pipeline as a high-level helper # Warning: Pipeline type "image-to-text" is no longer supported in transformers v5. # You must load the model directly (see below) or downgrade to v4.x with: # pip install "transformers<5.0.0" from transformers import pipeline pipe = pipeline("image-to-text", model="lightonai/LightOnOCR-1B-1025")# pip install -U transformers accelerate # Load model directly from transformers import AutoProcessor, AutoModelForSeq2SeqLM processor = AutoProcessor.from_pretrained("lightonai/LightOnOCR-1B-1025") model = AutoModelForSeq2SeqLM.from_pretrained("lightonai/LightOnOCR-1B-1025", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Update README.md
Browse files
README.md
CHANGED
|
@@ -67,18 +67,15 @@ All benchmarks evaluated using **vLLM** on the Olmo-Bench.
|
|
| 67 |
|
| 68 |
## Installation
|
| 69 |
|
|
|
|
| 70 |
```bash
|
| 71 |
|
| 72 |
uv venv --python 3.12 --seed
|
| 73 |
source .venv/bin/activate
|
| 74 |
|
| 75 |
-
|
| 76 |
-
uv pip install vllm
|
| 77 |
-
|
| 78 |
-
--torch-backend=auto \
|
| 79 |
-
--extra-index-url https://wheels.vllm.ai/${VLLM_COMMIT} \
|
| 80 |
-
--prerelease=allow
|
| 81 |
-
|
| 82 |
uv pip install pypdfium2 pillow requests
|
| 83 |
```
|
| 84 |
|
|
|
|
| 67 |
|
| 68 |
## Installation
|
| 69 |
|
| 70 |
+
[2025/11/24] 🚀 LightOnOCR is now officially supported in vLLM v0.11.1 🚀
|
| 71 |
```bash
|
| 72 |
|
| 73 |
uv venv --python 3.12 --seed
|
| 74 |
source .venv/bin/activate
|
| 75 |
|
| 76 |
+
# install any version higher than 0.11.1
|
| 77 |
+
uv pip install vllm==0.11.2
|
| 78 |
+
# extra deps need only to run the example below
|
|
|
|
|
|
|
|
|
|
|
|
|
| 79 |
uv pip install pypdfium2 pillow requests
|
| 80 |
```
|
| 81 |
|